Pollinator presence in orchards depends on landscape-scale habitats more than in-field flower resources

Abstract

Pollination is a critical ecosystem service given its essential role in sustaining food production, while pollinating insects are declining worldwide. Pollination capacity can be estimated through direct indicators characterising pollinator communities. Pollinators need feeding and nesting resources, so their presence can also be estimated through indirect indicators characterising these resources at plot and landscape scales. In this study, we aimed to identify the subset of resource indicators accounting for pollination capacity in orchards by relating insect presence to the most relevant environmental variables for managing pollinator presence. In 31 orchards of the Grenoble region (France) we measured direct indicators of pollinator abundance and richness at plot scale. Simultaneously we quantified indirect indicators of landscape and plot scale feeding and nesting resources. We selected indicators significantly correlated with insect presence using simple linear models between resource indicators and measures of insect taxonomic richness and abundance. Multiple linear regressions including significant resource indicators at both plot and landscape scales showed that landscape composition and presence of beehives explained between 19 and 63 % of the observed variance of pollinator community indicators. Total pollinator abundance decreased with distance to the closest grassland patch and increased with abundance of beehives. Pollinator richness increased with grassland cover within a 3 km radius. Domestic honeybee abundance increased with beehives located both in the plot and the landscape, and decreased with isolation from the closest grassland patch. Wild hymenopteran abundance increased with grassland cover within 3 km and forest cover at 500 m. Dipteran abundance increased when beehives were located only in the landscape. This study highlights the prevalent role of landscape-scale resources for pollinator communities over in-field flower resources. Thus, in order to favour pollinators, management should focus mainly on the landscape scale, necessitating cooperation between farmers. However, as resource indicators explained less than half of the observed variance for most community variables, inferring pollinator community characteristics from specific environmental variables remains uncertain. Further work could explore how orchard management practices influence our conclusions.